1) Today, the United States votes on its next president. 
2) NOT INVESTMENT ADVICE. NOT POLITICAL ADVICE.
3) Right now, prediction markets are around 38% for Trump. This is up from ~33% a week ago.
ftx.com/president2020
ftx.com/president2020
4) Polls have Biden ahead by roughly 8% in the national popular vote; @FiveThirtyEight has Trump at about 10% to win the electoral college.
projects.fivethirtyeight.com/2020-election-…
projects.fivethirtyeight.com/2020-election-…
5) US election results do not all come at once: instead each state, or sometimes district, reports separately.
The first significant returns start at 7pm EST = 8am HKT; Georgia is likely to be the closes state reporting then.
Results will trickle in over the proceeding 12h.
The first significant returns start at 7pm EST = 8am HKT; Georgia is likely to be the closes state reporting then.
Results will trickle in over the proceeding 12h.
6) Some states may take days to call, especially as they work to process mail-in-votes; others will be called the second the first results come in.
7) This may be one of the ugliest election nights.
There is a lot of arguing over votes, ballots, and procedures, and small changes can have large impact.
Something that disqualifies 5% of ballots in a single state could well swing the election.
There is a lot of arguing over votes, ballots, and procedures, and small changes can have large impact.
Something that disqualifies 5% of ballots in a single state could well swing the election.
8) Furthermore, Trump is signaling that he is unlikely to concede the election if it is at all contested, and may claim victory if it's close.
There is a real chance that two candidates will claim victory tonight.
That would be a shitshow.
There is a real chance that two candidates will claim victory tonight.
That would be a shitshow.
9) Prediction markets are running _way_ above models.
There are basically three theories here:
a) models don't understand how volatile the race is, or are missing Trump voters' enthusiasm
b) people are buying way too much Trump on prediction markets
c) contested elections
There are basically three theories here:
a) models don't understand how volatile the race is, or are missing Trump voters' enthusiasm
b) people are buying way too much Trump on prediction markets
c) contested elections
10) In particular it's worth noting that in the case of "extra-constitutional" events, what the models are predicting (electoral college vote) might settle differently than prediction markets (who is actually president).
11) This has probably been the most expensive election in history, with roughly $10b spent collectively.
The election is likely to dominate headlines for the next few days.
The election is likely to dominate headlines for the next few days.
12) Anyway we'll certainly be watching results closely, at fivethirtyeight.com and electionbettingodds.com.
(And, of course, ftx.com/president2020!)
(And, of course, ftx.com/president2020!)
13) Finally -- the first results have in fact already come in.
One small town in New Hampshire votes nearly a day before the rest of the world.
One small town in New Hampshire votes nearly a day before the rest of the world.
https://twitter.com/kylegriffin1/status/1323494113098960898
14) For more info on FTX election contracts, see help.ftx.com/hc/en-us/artic…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
